Do we need to talk about OpenWrt Rules?

[Rich Brown]

Hi,

> I fail to see what is wrong with current state.
> 
>> Should we have a six-month release schedule?
> 
> We've decided, but as you can see, didn't executed yet :-)

And this is exactly the problem I see. 

Because of our distributed decision-making process, no one (no single person) has the ability to decide when to pull the trigger on a release. So if we really do believe we need to make decisions, I believe we need a mechanism to make this explicit.

For example, we have decided on a six-month schedule and choose dates (say, October and April). All the features that are ready to go for the upcoming release would be included. Features that won't be ready will be held for the next release.

This is where the decision making process becomes important. For each release, we propose:

	"The following list of features will be in OpenWrt XX.YY to be released on ddMMMyyyy"

Then there's a discussion. We examine all the ramifications, decide whether we believe each feature can be complete by the specified date, maybe decide to postpone one of the features. (Please don't get hung up on time limits here - this discussion should go on as long as necessary.)

At some point, we'll have a sense of what's possible. We modify the proposal, then we vote. (This can have a fairly short timeline - we've already hashed out the issues.)

This gives clarity for everyone: people who promised that a feature will be ready have a firm deadline; people whose feature won't make it in that release can refine it ahead of the next release, or work on something else, or take a nap.

And it's always on-topic to ask for a change: For example, to start another discussion saying, "Feature X is taking a long time. Should we delay XX.YY by a week or two, or should we move it to the following release?"

Thank you again for this thoughtful discourse.

Rich